Chapter 4: Collecting Data

Question 1

Explanation for an outcome

Answer 1

1. by chance
2. discrimination

–> run simulation to find convincing evidence

Question 2

Define Population

Answer 2

entire group of your interest

Question 3

Define Sample

Answer 3

subset

Question 4

Define Census

Answer 4

Collects data from the entire population

Question 5

Types of Bad Sample

Answer 5

1. Convenience samples
2. Bias
3. Voluntary Response Sample

Question 6

Define Convenience Samples

Answer 6

• over/underestimate what you want to find from the population
• introduces bias
• produces samples that don’t reflect the population

Question 7

Define Bias

Answer 7

• systematically favoring a certain outcome
• when a study very likely to underestimate/overestimate what is being looked at

Question 8

Define Voluntary Response Sample

Answer 8

• made up of people who choose to answer a general appeal
• usually people with strong emotions

Question 9

Types of Good Sample

Answer 9

1. Simple Random Sample
2. Stratified Random Sample
3. Cluster Sample
4. Systematic Random Sample

Question 10

Define Simple Random Sample

Answer 10

An SRS of size __n__ is chosen so that every __GROUP__ of __n__ individuals has an equal chance to be selected as the sample

You must
1. numerically label the population
2. use technology to random digit table to get random numbers

Question 11

Define Sampling WITH replacement

Answer 11

individual can be selected more than once. repeats allowed

Question 12

Define Sampling WITHOUT replacement

Answer 12

individual cannot be selected more than once. repeats are ignored and not part of the sample.

Question 13

How to select an SRS

Answer 13

put everything in a pile and pick randomly
- label [ex. 001-100, ignore 000 and 101-999]
- use random number generator [RandInt(1,100,# that you want)] to select # and context
- no repeats

Question 14

Define Stratified Random Sample

Answer 14

divide population into strata (homogeneous group) and chose an SRS from each group and combine.
- more precise

Question 15

Difference between Simple and Stratified Random Sample

Answer 15

Simple: Every individual in the population has an equal chance of being selected

Stratified: the population is divided into groups (strata) based on characteristic, and then a random sample is taken from each stratum

Question 16

How to selected a stratified random sample

Answer 16

explain your choice of strata
- explain the strata [ex. strata can be types of books]
- randomly select __same number__ of __context__ from each __strata context__

“Not every GROUP of the same size has the same chance of being picked”

Question 17

Define Cluster Sample

Answer 17

create clusters (group) that “are located near each other”

randomly select a few clusters and include each member of the cluster.
- saves money and time

Question 18

Difference between cluster and stratified random sample

Answer 18

cluster: the groups are heterogeneous since they are physically located near
stratified: you don’t select from every cluster

Question 19

How to select a systematic random sample

Answer 19

1. calculation: (total amount)/(amount you want) = #
2. pick a number randomly from 1 - # and add # each time.

ex. arrange all books in order. Randomly select 1 book from the first 40. Then choose every 40th after that.

Question 20

Define Systematic Random Sample

Answer 20

Your population is somehow ordered

randomly select one of the first k individuals and choosing every kth individual
- good when population is ordered
- easier to conduct

Question 21

What can go wrong

Answer 21

1. undercoverage [members of the population have less of a chance of being chosen or left out]
2. nonresponse [chosen individuals can’t be contacted or refuse to participate –> big issue]
3. response bias [individuals lie or answer a question they don’t know]
4. question wording bias [the way a question is worded or asked influences the response from an individual]

Question 22

Define Experiment

Answer 22

has treatment
need experiment to know CAUSATION

Question 23

Define Observational Study

Answer 23

No treatment
need observational study to know correlation

Question 24

Define Confounding Variable

Answer 24

other possible variables other than explanatory variable that affects response variable

Question 25

Define Retrospective and Prospective

Answer 25

Retrospective = using existing data Prospective = tracks individuals into the future

Question 26

Define Treatment

Answer 26

specific condition applied to the experimental unit

Question 27

Define Experimental Unit/Subject

Answer 27

Person or thing treatment is randomly assigned to

Question 28

Define Factor

Answer 28

Explanatory variable that is manipulated and may cause. change int the response variable.

Question 29

Define Level

Answer 29

Different values of a factor

Question 30

4 Principles of Experimental Design

Answer 30

1. comparison 2. random assignment 3. control 4. replication

Question 31

Define Placebo Effect

Answer 31

when an individual responds to the dummy treatment

Question 32

Purpose of Control Group

Answer 32

provides a baseline for comparison

Question 33

Define single/double blind and their importance

Answer 33

Single-blind experiment = experimenters or subjects don't know what treatment they are getting or giving Double-blind experiment = both the subject and the experimenter do not know what treatment is given importance = blindness is important because it eliminates favoritism and/or placebo effect

Question 34

Purpose of Random Assignment

Answer 34

1. create roughly equivalent groups 2. helps control confounding variables

Question 35

How to randomly assign treatments to units

Answer 35

chance (not random sampling)

Question 36

Define Replication

Answer 36

use enough subjects/experimental units so the outcome of the experiment can have meaning

Question 37

Define Statistically Significant

Answer 37

the results most likely did not happen by chance convincing evidence (ran simulation)

Question 38

Define Blocking

Answer 38

like stratified sample. experiment version. benefits: 1. controls confounding variables 2. increases chance of finding convincing evidence if the effect is real

Question 39

How to determine to block

Answer 39

blocks should be homogeneous use a confounding variable

Question 40

Define Matched Pairs Design

Answer 40

blocks only include 2 experimental units 1. 2 similar units 2. 1 unit that gets both treatment --> randomly assign treatments or randomly assign the order of treatment

Question 41

Define Inference

Answer 41

using information from out sample/experiment to draw conclusions about the population

Question 42

Define Sampling Variability

Answer 42

different samples from the same population will give different results

Question 43

Define Margin of Error

Answer 43

"wiggle room" how far off our estimates is to the truth

Question 44

completely randomized design for experiments

Answer 44

hexagon shaped diagram. LOOK AT THE NOTES